Anthropocentric biocybernetic computing uses machine learning to provide a heightened level of understanding of human perceptions of complex systems. In this chapter, such processes are the catalyst for the development of two new methods for architectural and urban analysis and the revision and expansion of a third, existing method. The two new methods rely on, respectively, the 'Hough transform' algorithm to model line segmentation in images of buildings and a newly authored program to investigate facial pareidolia in façades. The third method is a computational variation of the fractal analysis technique for the classification and review of urban skylines. Anthropocentric biocybernetic methods are ideal for the analysis of the built environment because architecture evokes a complex emotional and perceptual response in viewers and building users. While architectural and urban spaces may be readily measured and understood in terms of their formal characteristics, the phenomenal and semiotic qualities of the built environment have rarely been investigated using computational means. This chapter responds to this situation by providing examples of new techniques that are ideal for expanding the computational analysis of architecture beyond the simple analysis of form.
Built Environment: Design Management and Applications p. 121-145